W e have evaluated the possibility of inducing antibody responses locally in the human stomach as a prerequisite for the development of a vaccine against Helicobacter pylori. Both H. pylori-infected and noninfected subjects were immunized with an oral B subunit whole cell (BS-WC) cholera vaccine, and total and vaccine-specific antibody-secreting cells (ASC) were determined by the enzyme-linked immunospot (ELISPOT) technique in cells isolated from the antrum and duodenum, respectively, before and after vaccination. Most of the subjects responded to the vaccination with high frequencies of vaccine-specific ASCs in the duodenum as well as high-serum antibody titers, and no significant differences were seen in the responses between H. pylori- infected and noninfected subjects. When studying the gastric mucosa, on the other hand, there were dramatic differences between the H. pylori-infected and the noninfected subjects. Thus, whereas none of the noninfected subjects responded to the immunization in antrum, most of the H. pylori-infected subjects had high frequencies of vaccine-specific ASCs in this location after vaccination. Furthermore, the H. pylori-infected subjects had strikingly higher (as a mean 80-fold) frequencies of total IgA-secreting cells in antrum than the noninfected subjects, whereas the frequencies of total IgA-secreting cells in the duodenum were comparable between the groups. In conclusion, these results demonstrate the possibility of inducing antibody responses locally in the gastric mucosa of H. pylori-infected individuals, a finding with obvious implications for the future development of a therapeutic vaccine against H. pylori.